tokens-per-dollar · prices read
Copilot AI credits to dollars
One GitHub AI credit is $0.01. Copilot Pro ($10 a month) includes $15 of credits, Pro+ ($39) $70 and Max ($100) $200. Of those, $5, $31 and $100 are a flex allotment GitHub says may change over time. Credits are spent per token, at a rate per model: on Claude Opus 5.5, $4.00 per million input tokens and $20.00 per million output, so one credit buys 2,500 input or 500 output tokens. Past the included credits, usage is billed at the same rates. Copilot keeps no logs on your machine, so the closest you can get to your own month is your Claude Code and Codex tokens at these rates, which npx --allow-git=root github:agentwares/tokens-per-dollar works out and labels an estimate.
What each plan includes
| Plan | Per month | AI credits included | Of which flex | Credits per dollar of fee |
|---|---|---|---|---|
| Copilot Pro | $10 | 1,500 ($15) | $5 | 150 |
| Copilot Pro+ | $39 | 7,000 ($70) | $31 | 179 |
| Copilot Max | $100 | 20,000 ($200) | $100 | 200 |
Credits reset on the 1st of the month (UTC) and do not roll over. Code completions are not billed in credits.
Tokens per credit, by model
| Model | Per million: input · cached · cache write · output | Tokens per credit: input · output |
|---|---|---|
| Claude Fable 5.1 | $10.00 · $0.25 · $12.50 · $50.00 | 1,000 · 200 |
| Claude Opus 5.5 | $4.00 · $0.20 · $5.00 · $20.00 | 2,500 · 500 |
| Claude Sonnet 5.5 | $2.00 · $0.10 · $2.50 · $10.00 | 5,000 · 1,000 |
| Claude Haiku 4.5 | $1.00 · $0.10 · $1.25 · $5.00 | 10,000 · 2,000 |
| GPT-6 Astra | $10.00 · $1.00 · $12.50 · $50.00 | 1,000 · 200 |
| GPT-6.1 Sol | $2.00 · $0.10 · $2.50 · $10.00 | 5,000 · 1,000 |
| GPT-6 Luna | $0.10 · $0.01 · $0.125 · $0.50 | 100,000 · 20,000 |
| GPT-5.6 Sol | $4.00 · $0.40 · $5.00 · $20.00 | 2,500 · 500 |
GitHub's standard rates. Several models cost more above a prompt size (272,000 tokens for most GPT models), and tokens-per-dollar applies those rows too. GitHub lists one cache-write price per Claude model, where Anthropic's API charges more for writes to the 1-hour cache, so the same Claude Code tokens can come to slightly less on Copilot than on the API.
Why your month is an estimate here
Claude Code and Codex write every response's token counts to disk; Copilot does not, and its agent would not use exactly the tokens another agent used for the same work. So the Copilot rows in npx --allow-git=root github:agentwares/tokens-per-dollar answer a narrower question, “what would these tokens cost on Copilot”, and say so on every row.
Changelog: what the vendors changed, and when
- · GitHub: Every Copilot plan billed in GitHub AI credits, each token priced by model, replacing premium requests (which remain only on legacy annual plans). GitHub changelog, 1 June 2026: updates to Copilot billing and plans
Dates are the vendors' own. “As reported” means the press quoting the vendor, because the vendor's own post could not be read. Each entry says what changed, never why.
Run it on your own logs
On the machine where you use Claude Code or Codex:
$ npx --allow-git=root github:agentwares/tokens-per-dollarIt reads the transcripts Claude Code and Codex keep on your machine, takes the token counts, and prices your last 30 days on every plan on this page. Free and MIT; it makes no network request and sends nothing. --json prints every price with its source and the date it was read.
For a team: seatledger keeps everyone's coding-agent ledger in one place and flags when the token rate moves. Free for up to 3 developers; email alerts on Team, $49 a month.
Sources
- GitHub Copilot plans — read 8 October 2026
- Copilot billing for individuals (1 AI credit = $0.01) — read 8 October 2026
- Copilot models and pricing (per-token rates) — read 8 October 2026
- GitHub changelog, 1 June 2026: updates to Copilot billing and plans — read 8 October 2026